摘要
根据单目视觉高精度位姿测量中合作目标形状和图案设计的基本原则,分析了在透视投影情况下方形和圆形合作目标特征点的提取精度,指出在特征点提取时方形合作目标角点具有透视不变性、圆形合作目标圆心存在固有偏移误差的特点,并得出减小单个圆形目标尺寸容易造成抗污染能力差而方形目标抗污染能力强的结论。针对光阴影干扰测量精度的问题,提出将方形合作目标外边缘作为测量基准、内边缘斜率作为修正因子的角点提取算法。基于单目视觉位姿测量系统内部和外部参数,给出了合作目标几何尺寸设计和加工精度要求的计算方法。最后,通过合作目标设计实例和实验,验证了方形合作目标能满足复杂光干扰环境下高精度位姿测量需求。
Based on the basic principle of shape and pattern design of cooperative targets in monocular vision high precision pose measurement,the extraction accuracy of the feature points of square and circular cooperative targets in perspective projection is analyzed in this study.The characteristics of the corner points of square cooperative targets exhibiting perspective invariance and the center of circular cooperative targets exhibiting an inherent deviation error are identified when feature points are extracted,concluding that reducing the size of a single circular target can result in poor anti-pollution but strong anti-pollution ability of a square target.Aiming at addressing the problem of the measuring accuracy of light shadow interference,a corner extraction algorithm using the outer side as the measurement reference and the inner side as the correction factor of square cooperative targets is proposed herein.Furthermore,based on the internal and external parameters of a monocular visual pose measurement system,the calculation method for geometric dimension design and machining precision of cooperative targets is presented.Finally,through the design sampling and experimentation of cooperative targets,it is verified that the square cooperative targets can meet the demands of high-precision pose measurement in a complex optical interference environment.
作者
屈也频
刘坚强
侯旺
Qu Yepin;Liu Jianqiang;Hou Wang(Naval Research Academy,Shanghai 200436,China)
出处
《光学学报》
EI
CAS
CSCD
北大核心
2020年第13期118-125,共8页
Acta Optica Sinica
基金
“十三五”装备预先研究项目(3020901020101)。
关键词
机器视觉
单目视觉测量
合作目标
图形设计
亚像素精度
阴影补偿
透视不变性
machine vision
monocular vision measurement
cooperative target
graphic design
sub-pixel precision
shadow compensation
perspective invariance